PdM Common Pitfalls

icon-Support

Predictive Maintenance - Common pitfalls and challenges

As you would expect the common pitfalls for a PdM project are closely related to the best practices. Encountering any of the challenges outlined below can lead to your Preventative Maintenance programme failing to achieve its aims or delivering optimum value for your organisation.

When selecting and using equipment and systems for Predictive Maintenance it is essential for everyone involved to understand fully what the overall purpose is for that component. Not properly understanding why a component or piece of technology is to be used will lead to ineffective and incorrect usage and failure to deliver the overall programme aims.

The work required across the programme to integrate and enable data to flow in a timely manner through systems across your key processes e.g. ERP, MES, CMMS, and Data Platform should not be underestimated. Ensure that an overall Integration Strategy is developed and that ease of integration is one of the key selection criteria for any component being implemented by the programme.

When selecting a team for a PdM programme always look for the right skills from both internal and external resources.

The best candidates may not be the obvious ones. Implementing Predictive Maintenance will require a different approach to more basic maintenance approaches. Having a team that is open to change; excellent at problem solving; analytical in nature; flexible; willing to learn; inquisitive; and communicates well will be a significant factor in the success or not of a PdM programme.

For external resources make sure they can evidence their experience, expertise, and (ideally) certification in the areas you are seeking them to undertake.

Training is a critical requirement to enable a successful PdM implementation. All the personnel involved need to be properly trained in the new processes and technologies they will be using. They should also understand and be supportive of the overall vision for the programme. This is one of the most common failures that impact PdM programmes.

Both initial and on-going training will be required to ensure the required skills are established and developed over time.

Moving to a Predictive Maintenance strategy (moving from fix on failure or preventive maintenance strategy) will require a shift in mindset across the organisation. Ensure that sufficient time and effort is planned for and deployed to enable this to happen. The programme will need to be dynamic and able to flex over time. Actively look for ways to keep the programme fresh by finding new applications and opportunities to increase equipment reliability.

Rolling out Predictive Maintenance across an organisation requires consistency in all the methods, procedures, and processes adopted in order to maximise the benefits realised. There are many areas where a lack of consistency and thus an absence of repeatability can cause a PdM programme to fall down. For example, if the accuracy of the data being collected is compromised then the accuracy of the analysis and recommendations on machine health made will be affected thus impacting the corrective actions that may or may not be taken. 

At the heart of a PdM implementation is gathering and analysing the required data sets in a consistent and timely manner. Not managing this frequently leads to failure. Those involved in the data gathering processes need to understand the importance of their role in the overall implementation and be held accountable for it.

Some organisations believe they don’t have enough data to even get started with Predictive Maintenance. This can be addressed by e.g. analyse what you have (there may be more to it than expected); changing the data capture mechanisms to capture more data; using simulation tools to generate synthetic data.

Predictive analysis requires data classified as success or failure and needs sufficient volume of each type to successfully train models. If there is insufficient failure data you will need to synthesise some. Fully understanding the warning signs from patterns within the operational (success) data will help inform the characteristics needed in the failure data.

Having active sponsorship for the whole programme lifecycle through to business as usual adoption and beyond is critical for a successful implementation. As a minimum, the sponsor(s) should be a vocal advocate for the programme with senior management and enable the allocation of the required funding and resources.

As a continuous improvement initiative regular, accurate measurement of outcomes is key alongside effective communication of the benefits realised is a key requirement for a PdM implementation. Identifying effective KPIs as early in the programme as possible and regularly reviewing their suitability is essential.  

Programmes often fail to define strong PdM KPIs and then only communicate initial programme wins. It is recommended that a programme scorecard is used to capture the numbers and types of wins throughout the programme lifespan. This can be used to ensure that costs savings and other benefits realised are accurately reported.

Predictive Maintenance is not just putting in place the mechanisms to gather and analyse the necessary machine data. Without taking the prescribed remedial actions the predicted equipment failures will still occur. Rolling out appropriate processes and training particularly for the planning and execution of follow-up activities is essential to PdM success.

There is more detail on how to avoid the common pitfalls for Predictive Maintenance programmes in the Resources section.